This page contains live results for Pennsylvania. My goal is not to predict the winner of the state. In fact, it’s the opposite. My goal is to understand where the tallied votes are from, where the untallied votes are from, and encourage caution in making assumptions on incomplete data.
[TEST MODE: These results are currently fake.]
There are three categories of ballot: in-person, mail-in that has already been received (“returned”), and mail-in that has not yet been received (“outstanding”). Any ballot postmarked by 11/3 will be counted.
*: It’s impossible in the PA data portal to tell how many of the returned mail-ins are included in the totals. The Live Results contain all of the in-person votes and an unknown fraction of returned mail-ins. In the Scenarios, I assume that zero mail-ins have been counted, to stay conservative about the range of possible eventual results.
In-Person Precincts Reporting: 9,127 / 9,128
Votes Tallied: 2,708,293
Biden: 1,249,440 (46.1%)
Trump: 1,043,311 (38.5%)
County-Weighted Change: +7.8% D
Returned Mail-Ins: 435,501
% Dem: 337,435 (77.5%)
% Rep: 63,760 (14.6%)
% Other: 34,306 (7.9%)
Outstanding Mail-Ins: 2,133,632
% Dem: 1,330,805 (62.4%)
% Rep: 560,947 (26.3%)
% Other: 241,880 (11.3%)
Data Sources
Live Results from electionreturns.pa.gov. (Updated 2020-10-15 08:44:36)
Mail-In data from data.pa.gov. (Updated 2020-10-14 21:29:33)
Scenarios
Final outcomes under the following assumptions…
If mail-ins vote by party, and 20% of the outstanding mail-ins are eventually returned.
Biden: 1,894,377 (53.1%)
Trump: 1,260,601 (35.3%)
If mail-ins vote by party, and no more outstanding mail-ins are eventually returned.
Biden: 1,604,028 (51%)
Trump: 1,124,224 (35.8%)
If mail-ins vote like current results adjusted for party[1], and 20% of the outstanding mail-ins are eventually returned.
Biden: 1,834,624 (58.4%)
Trump: 1,151,693 (36.6%)
If mail-ins vote like current results adjusted for party[1], and no more outstanding mail-ins are eventually returned.
Biden: 1,552,902 (49.4%)
Trump: 1,090,649 (34.7%)
[1] - I regress current county-level results on party registration. Then apportion mail-ins using those results. This assumes that in-person voters are representative of registered voters and in-person Democrats vote like mail-in Democrats (and Republicans like Republicans). This may or may not be valid.